Alpha/beta interferons regulate murine gammaherpesvirus latent gene expression and reactivation from latency

Abstract

Alpha/beta interferon (IFN-alpha/beta) protects the host from virus infection by inhibition of lytic virus replication in infected cells and modulation of the antiviral cell-mediated immune response. To determine whether IFN-alpha/beta also modulates the virus-host interaction during latent virus infection, we infected mice lacking the IFN-alpha/beta receptor (IFN-alpha/betaR(-/-)) and wild-type (wt; 129S2/SvPas) mice with murine gammaherpesvirus 68 (gammaHV68), a lymphotropic gamma-2-herpesvirus that establishes latent infection in B cells, macrophages, and dendritic cells. IFN-alpha/betaR(-/-) mice cleared low-dose intranasal gammaHV68 infection with wt kinetics and harbored essentially wt frequencies of latently infected cells in both peritoneum and spleen by 28 days postinfection. However, latent virus in peritoneal cells and splenocytes from IFN-alpha/betaR(-/-) mice reactivated ex vivo with >40-fold- and 5-fold-enhanced efficiency, respectively, compared to wt cells. Depletion of IFN-alpha/beta from wt mice during viral latency also significantly increased viral reactivation, demonstrating an antiviral function of IFN-alpha/beta during latency. Viral reactivation efficiency was temporally regulated in both wt and IFN-alpha/betaR(-/-) mice. The mechanism of IFN-alpha/betaR action was distinct from that of IFN-gammaR, since IFN-alpha/betaR(-/-) mice did not display persistent virus replication in vivo. Analysis of viral latent gene expression in vivo demonstrated specific upregulation of the latency-associated gene M2, which is required for efficient reactivation from latency, in IFN-alpha/betaR(-/-) splenocytes. These data demonstrate that an IFN-alpha/beta-induced pathway regulates gammaHV68 gene expression patterns during latent viral infection in vivo and that IFN-alpha/beta plays a critical role in inhibiting viral reactivation during latency.

FIG. 1.

Lethality and viral replication following γHV68 infection in mice lacking IFN signaling pathways. (A) Age-matched 129SvPas, IFN-α/βR^−/−, IFN-γR^−/−, IFN-α/βγR^−/−, or STAT1^−/− mice were infected with 100 PFU γHV68 intranasally and observed daily for mortality. Numbers of mice infected for each genotype are indicated, and results are compiled from three to five independent experiments. (B) Age-matched 129SvPas or IFN-α/βR^−/− mice were infected with 100 PFU γHV68 intranasally, and spleen, liver, and lung were harvested 7 to 9 or 21 to 28 dpi. Live virus present in each organ was quantitated by plaque assay. Shown are viral titers present in individual organs from six to nine mice in three independent experiments.

FIG. 2.

Establishment of viral latency in wt and IFN-α/βR^−/− mice. Splenocytes and peritoneal cells from 129SvPas and IFN-α/βR^−/− mice were harvested 28 dpi, and the frequency of viral genome-bearing cells was determined using limiting dilution nested PCR for the viral ORF72 gene. Shown are means and standard errors of the means pooled from three independent experiments. Each sample contained pooled cells from three to five mice. Dashed lines indicate the point of 63% Poisson distribution, determined by nonlinear regression, which was used to calculate the frequency of cells bearing viral genome.

FIG. 3.

Reactivation from latency and lytic virus persistence in wt, IFN-α/βR^−/−, and IFN-γR^−/− mice. Splenocytes (A and C) and peritoneal cells (B and D) from wt, IFN-α/βR^−/−, and IFN-γR^−/− mice were harvested 28 dpi, and the frequency of cells capable of reactivating lytic virus was determined by 21-day culture on indicator fibroblasts. Wells were scored as positive for viral reactivation based on the presence of complete CPE by microscopic observation (A and B). Parallel samples were mechanically disrupted prior to culture to detect preformed lytic (persistent) virus (C and D). Shown are means and standard errors of the means pooled from three to four independent experiments. Each sample contained pooled cells from three to five mice. Dashed lines indicate the point of 63% Poisson distribution, determined by nonlinear regression, which was used to calculate the frequency of cells reactivating lytic replication.

FIG. 4.

Effects of IFN-α/βR on viral reactivation wane with time. (A) Peritoneal cells from wt or IFN-α/βR^−/− mice were harvested 161 dpi, and the frequency of cells capable of reactivating lytic virus was determined by 21-day culture on indicator fibroblasts as in Fig. 3. (B) Parallel samples were mechanically disrupted prior to culture to detect preformed lytic (persistent) virus. Shown are means and standard errors of the means pooled from four independent experiments. Each sample contained pooled cells from three to five mice. Dashed lines indicate the point of 63% Poisson distribution, determined by nonlinear regression, which was used to calculate the frequency of cells reactivating lytic replication.

FIG. 5.

Depletion of IFN-α/β in vivo results in increased viral reactivation in wt mice. wt mice were infected with 100 PFU γHV68 intranasally. At 21 and 25 dpi, mice received 30 μl control sheep serum or IFN-α/β-depleting sheep serum (5, 33) intraperitoneally in a volume of 500 μl phosphate-buffered saline. Peritoneal cells were harvested 28 dpi, and viral reactivation was assessed on indicator fibroblasts in the presence of a 1:2,500 dilution of control or IFN-α/β-depleting serum. In a parallel experiment, addition of IFN-α/β-depleting serum to reactivation cultures was not sufficient to alter reactivation efficiency in the absence of in vivo depletion (data not shown). Shown are means and standard errors of the means for three individual mice in each group. Dashed lines indicate the point of 63% Poisson distribution, determined by nonlinear regression, which was used to calculate the frequency of cells reactivating lytic replication. The experiment was repeated once with comparable increases in reactivation in IFN-α/ β-depleted mice.

FIG. 6.

Effect of IFN-α/β of inhibiting γHV68 reactivation cannot be complemented in trans. Peritoneal cells from wt or IFN-α/βR^−/− mice were harvested 28 dpi, and the frequency of cells capable of reactivating lytic virus was determined by 21-day culture on indicator fibroblasts as in Fig. 3. Cells from wt and mutant mice were plated independently or mixed in equal numbers prior to serial dilution and plating on indicator fibroblasts. Shown are means and standard errors of the means pooled from two independent experiments. Each sample contained pooled cells from three to five mice.

FIG. 7.

Latent viral gene expression in wt and IFN-α/βR^−/− mice. Spleens were harvested from wt or IFN-α/βR^−/− mice 28 dpi, and total RNA was purified. (A) Twenty micrograms of total RNA was hybridized overnight with the γHV68-specific γ6 riboprobe set and digested with RNase, and protected fragments were resolved by polyacrylamide gel electrophoresis. Each lane represents pooled spleen RNA from two to three mice (experiment 1) or from individual mice (experiment 2). Two micrograms of total RNA from γHV68-infected owl monkey kidney cells was used as a positive control (+). Shown are phosphorimages from two independent experiments. (B) Viral and cellular transcript levels were quantified by phosphorimager analysis and are presented as phosphorimager (PI) units. The means and standard errors of the means for phosphorimager units are shown for each transcript following quantitation of phosphorimages in panel A.